Nazariy – Python, Machine learning, Cloud Architecture

Teaching "Recommender Systems" course to Master's students

• Created a modern course about recommender systems.

Socially responsible job platform

• Designed recommendation system, which doubled all of the core metrics
• Set up ML infrastructure and workflow
• Developed and integrated multiple SOTA NLP models

AI startup in content localization domain.

• Built Proofs-of-Concept for AI functionality
• Designed and executed an implementation strategy
• Established a cloud architecture and streamlined CI/CD processes
• Built and managed a highly-effective remote development team

Product for media monitoring and inauthentic behaviour detection

• Designed and built ML layer of the application, including models, modules and pipelines for text processing and narrative classification
• Set up cloud-based infrastructure
• Optimised application accuracy, efficiency, and performance

Chatbots are used in banking, healthcare, and retail with advanced NLP models: sentiment measurements, intent classification, entities extraction Chatbots are developed with optimal technologies depending on the client's ecosystem

• Developed complex automated data pipelines and conversation flows
• Created NLP models using TensorFlow and Pytorch
• Integrated 3rd party APIs and services
• Projected unified conversational bots across various communication channels: Slack, Microsoft Teams, Telegram, and Viber

ClipIt is a short-form, video-sharing app that allows gamers to create and share clips from their gaming experience

• Architected and designed Machine Learning infrastructure
• Developed recommendation engine
• Set up CI/CD pipelines in the cloud

Pfizer Inc is an American multinational pharmaceutical and biotechnology corporation headquartered on 42nd Street in Manhattan, New York City.

• Implemented and iteratively improved various NLP models
• Worked on topic modeling and search engine
• Worked on data optimization
• Integrated Machine Learning solutions in the back-end solution

Strategy for impact attribution across various marketing channels (Facebook, google analytics, etc.) The final model incorporated click-based, incrementality-based, and ML-based attribution approaches and allowed to optimize marketing spending.

• Designed software architecture
• Implemented the model with SQL and Python, and presented the results to stakeholders

Various products in gaming, education and fintech domains

• Communicated closely with business owners and rapidly prototyped solutions for business needs
• Built development teams and managed them during project delivery
• Ensured best development practices
• Participated in the migration from monolith to microservice architecture by designing a bundle of microservices
• Configurated CI/CD pipelines, which boosted development and delivery speeds
• Configurated complex pipelines for automatic code quality checks and integration/unit tests
• Analyzed and resolved cloud performance issues
• Designed uniform API and database versioning strategies